A theory that has gained traction in the renewable energy space is that renewable energy sources like wind and solar are based on manufactured “technologies”, while fossil fuel energy sources like oil, coal, and natural gas are based on extracted “commodities”.
The renewable=technology and fossil fuels=commodity argument seems more like a PR attempt to increase government subsidies for renewables. And the logic is backwards. If renewables will drastically decrease in cost regardless, then there is no need for government subsidies.
Note that population growth must also be considered. If population doubles, like it has since 1975, and prices remain the same, we’re effectively doubling economic supply.
The technology versus commodity dichotomy seems a bit arbitrary.
Renewables avoid fuel commodity, but still have great dependence on commodities on their manufacturing and certainly their low density necessitating long grid extension to move the power to where it is needed. They also suffer from an inability to stockpile necessitating commodity rich storage technologies.
Commodity fuel, of course, requires other commodities and technology to be converted into power.
To your point of commodities that were produced for thousands of years: I'm not sure "learning curves" are actually a meaningful concept prior to the industrial revolution, for a couple of reasons:
Learning curves require producers having high visibility over each other, competitive dynamics, a culture of growth/improvement, and reliable ways of embedding accrued technical knowledge generation on generation.
Iron may have been extracted and refined in China, Mali, and England, but it's not clear they learned anything from each other for most of that time.
The only manufactured good categories that experienced really significant declines in the AEI list are consumer electronics (computers et al), TV/AV equipment, and cameras. These 5 categories radically skew the 2000-2020 chart data. Eliminate them, and the peak of that chart gets recentered on 0%. The so-called "exponential growth curve of technologies" applies mostly to digital technologies.
So the question is: are windmills and solar panels like most other manufactured goods, or are they more like computers and TVs and smartphones?
To admit my own bias, I think windmills bear absolutely no resemblance to digital consumer electronics and do not expect their cost so exhibit double-digit drops like those do. Solar panels are a harder case, but I expect most of the production cost declines have already occurred.
I don't think this is true at all. The breakdown of CPI categories I uses includes lots of manufactured good that aren't consumer electronics or similar things, and only 9 items out of the 67 increased in inflation-adjusted price from 2000 to 2020.
Are you using different data for the chart then? Because the AEI-like chart only shows 7 categories with 7%+ declines in price over 2000-2019: comp, telecom, TVs, video, audio, cameras and lamps (I wonder if that one is due to the LED transition?)
Solar has been the rock star of RE, with the price of the actual flat plastic thingies that collect the sunlight having fallen by nine-tenths in the last decade or so. The flat plastic thingies have become a minor, almost insignificant part of the cost of a solar installation. A flat plastic thingy has a socket kind of thingy on it to connect it to the household circuit or to the national grid or whatever. From the socket thingy onward, costs have not come down by nine-tenths; they have risen in line with normal increases in the costs of materials and labor. Further cost reductions seem unlikely.
Natural gas has been the rock star of FF, with prices dropping dramatically after the introduction of horizontal drilling technology. Further cost reductions seem unlikely. Prices of both solar and natural gas have settled into a groove. Future gains will be incremental, not exponential.
Is light a particle or a wave, is solar a commodity or a technology, is fracking a commodity or a technology, as you say the lines have become blurred.
The real materials that are used to make solar and wind power plants and grid lines and batteries etc. are all finite, just like fossil fuels. And of course these technologies require copious amounts of fossil fuels to manufacture, from the mining and refining of the materials, to the factories, to installing, maintaining, etc. Dams, too, require a LOT of materials, and require constant maintenance. These are strict limits on a finite planet. Given the rapidly dwindling ore grade of many materials critical to these technologies, it seems likely costs will increase. More diesel, more machines, more man hours, more infrastructure, all required to extract and refine the materials to make these and all technologies ensures it. And of course eventually these materials will "run out" in the sense that it will take more energy or more money or both to extract and refine them than manufacturers can make back. Just like EROEI of fossil fuels is (for the most part) declining. These are the obvious and inevitable consequences of accelerating overshoot.
(This doesn't take into account, of course, the massive ecological footprint of these technologies, which is steadily increasing as renewables are rolled out. This too has a "cost" -- no living animal, not even humans, can survive without flourishing ecosystems.)
Just tagging things as commodities vs. technology seems unreasonably simplistic. Each source has its own analysis. It grates to classify resources as simply commodities - clothing is a commodity which is ridiculous to lump with iron. And iron is not even like oil or molybdenum or semi-grade silicon or LOX or lanthanides.
Each of these has their own story. You can analyze coal or oil or solar or wind, in particular, and there seems nothing to learn from the game of piling them all into "commodity" or "technology" and expecting a story to fit. Sure you might gain the occasional insight - but nothing comparable to a serious analysis. Oil vs. solar for example each involve resource extraction, land use, pollution, questions of renewability, replacement rate, etc. In each of these dimensions you can make some interesting comparisons or contrasts.
But an overall narrative of technology vs. commodity? At the end of your essay nothing really emerges except perhaps "don't take guidance from such a superficial urge to categorize".
If you want to understand where things are headed, there is plenty of detail worth studying.
Enjoyed the read. The real obstacle isn’t whether renewables “work,” it’s that oil is so damn cheap, and so deeply embedded in our society/processes, that it’s become the easiest way to keep a complex economy running...which makes moving away from it structurally hard, even when people want to move to renewables. So frustrating.
The distinction is irrelevant. All energy sources require technology to extract energy from the commodity. Fossil fuel and hydro power generation technologies have been around a long time, and aren't likely to get much more efficient at the point of generation. Nuclear, solar, and wind generation technologies still have room to improve generating efficiency. Fossil fuel and nuclear fuels still have scope to increase in extraction efficiency. (Wind, solar, and hydro generate power where they're extracted, so the distinction there doesn't really matter.) But the longer a power source has been in use, the smaller the individual gains will generally be because there have been more attempts to improve.
Distributed production isn't always a net benefit. Generating power in individual vehicles is generally less efficient than generating power centrally and storing it in the vehicle, but has significant convenience benefits. At the scale of the national grid, 5GW power stations are still a "distributed" source; no one failure will cripple the system. It's easier to harden a single power station against the weather than it is to protect a million solar panels. But more finely distributed generation is more robust to transmission network failures.
Lastly, short and medium them price fluctuations aren't entirely cause by changes in the technology of extraction or generation. All phases of energy production, distribution, and finance are subject to political interference; major price changes can be purely the result of changes in policy halfway around the world.
It is puzzling that you would use materials from RMI which is simply an activist organization and their reports are always the same. This difference between technologies and commodities is not very useful and is mainly a way to buttress renewables instead of turning things like solar panels into commodities that are priced below market pricing from China. Much more interesting and useful would be to look at the work of Alvin Roth the recent winner of the Nobel in Economics who makes a much more meaningful distinction between commodity markets and matching markets and how to apply that thinking properly to the construction industry.
What if instead of "technology vs commodity" we expanded the analysis into "manufacture vs construction vs harvested vs extracted"? I feel like your book/work has shown that there are better learning curves for manufactures vs construction, and as you're showing now there appears to be some difference between agriculture and mining. I just mean "technology vs commodity" might be too simplistic, especially since as you point out, "technology" is kind of everything. But then it would still get to the point of "can things move from extraction to manufacture?" "Can things move from construction to manufacture?" How much does it help, and/or where does it break down? (Hydropower is construction, but if you get better at manufacturing them you still run out of places to put them. Windmills are manufactures but wake dynamics limit the useful supply. Etc.)
The simulations of production-curves-with-missing-data are showing >8 orders of magnitude on the x-axis: that seems unrealistic for looking at real-world datasets. Obviously someone somewhere collected the first bushel of grain, or mined the first kilogram of iron ore, but you wouldn't know its price; the best you'd know would be the annual production totals. For something like crops, those annual totals should grow pretty slowly over the 100-year stretch; e.g. for US crops, total annual production in 1900 is presumably >10% of what it is today (because population was ~1/3x), which means that the cumulative volume over that period should only span 3-4 orders of magnitude. It might be worth trying to simulate more realistic estimates of which x points you would expect to see.
Interesting article.
The renewable=technology and fossil fuels=commodity argument seems more like a PR attempt to increase government subsidies for renewables. And the logic is backwards. If renewables will drastically decrease in cost regardless, then there is no need for government subsidies.
Another excellent essay.
Note that population growth must also be considered. If population doubles, like it has since 1975, and prices remain the same, we’re effectively doubling economic supply.
The technology versus commodity dichotomy seems a bit arbitrary.
Renewables avoid fuel commodity, but still have great dependence on commodities on their manufacturing and certainly their low density necessitating long grid extension to move the power to where it is needed. They also suffer from an inability to stockpile necessitating commodity rich storage technologies.
Commodity fuel, of course, requires other commodities and technology to be converted into power.
To your point of commodities that were produced for thousands of years: I'm not sure "learning curves" are actually a meaningful concept prior to the industrial revolution, for a couple of reasons:
Learning curves require producers having high visibility over each other, competitive dynamics, a culture of growth/improvement, and reliable ways of embedding accrued technical knowledge generation on generation.
Iron may have been extracted and refined in China, Mali, and England, but it's not clear they learned anything from each other for most of that time.
The only manufactured good categories that experienced really significant declines in the AEI list are consumer electronics (computers et al), TV/AV equipment, and cameras. These 5 categories radically skew the 2000-2020 chart data. Eliminate them, and the peak of that chart gets recentered on 0%. The so-called "exponential growth curve of technologies" applies mostly to digital technologies.
So the question is: are windmills and solar panels like most other manufactured goods, or are they more like computers and TVs and smartphones?
To admit my own bias, I think windmills bear absolutely no resemblance to digital consumer electronics and do not expect their cost so exhibit double-digit drops like those do. Solar panels are a harder case, but I expect most of the production cost declines have already occurred.
I don't think this is true at all. The breakdown of CPI categories I uses includes lots of manufactured good that aren't consumer electronics or similar things, and only 9 items out of the 67 increased in inflation-adjusted price from 2000 to 2020.
Are you using different data for the chart then? Because the AEI-like chart only shows 7 categories with 7%+ declines in price over 2000-2019: comp, telecom, TVs, video, audio, cameras and lamps (I wonder if that one is due to the LED transition?)
So what am I missing?
Solar has been the rock star of RE, with the price of the actual flat plastic thingies that collect the sunlight having fallen by nine-tenths in the last decade or so. The flat plastic thingies have become a minor, almost insignificant part of the cost of a solar installation. A flat plastic thingy has a socket kind of thingy on it to connect it to the household circuit or to the national grid or whatever. From the socket thingy onward, costs have not come down by nine-tenths; they have risen in line with normal increases in the costs of materials and labor. Further cost reductions seem unlikely.
Natural gas has been the rock star of FF, with prices dropping dramatically after the introduction of horizontal drilling technology. Further cost reductions seem unlikely. Prices of both solar and natural gas have settled into a groove. Future gains will be incremental, not exponential.
Is light a particle or a wave, is solar a commodity or a technology, is fracking a commodity or a technology, as you say the lines have become blurred.
The real materials that are used to make solar and wind power plants and grid lines and batteries etc. are all finite, just like fossil fuels. And of course these technologies require copious amounts of fossil fuels to manufacture, from the mining and refining of the materials, to the factories, to installing, maintaining, etc. Dams, too, require a LOT of materials, and require constant maintenance. These are strict limits on a finite planet. Given the rapidly dwindling ore grade of many materials critical to these technologies, it seems likely costs will increase. More diesel, more machines, more man hours, more infrastructure, all required to extract and refine the materials to make these and all technologies ensures it. And of course eventually these materials will "run out" in the sense that it will take more energy or more money or both to extract and refine them than manufacturers can make back. Just like EROEI of fossil fuels is (for the most part) declining. These are the obvious and inevitable consequences of accelerating overshoot.
(This doesn't take into account, of course, the massive ecological footprint of these technologies, which is steadily increasing as renewables are rolled out. This too has a "cost" -- no living animal, not even humans, can survive without flourishing ecosystems.)
The graphs are to die for. Thanks for sharing, Brian!
event technologies such as batteries, at scale, require huge labour, capital and expertise cost that might make it less obvious
Just tagging things as commodities vs. technology seems unreasonably simplistic. Each source has its own analysis. It grates to classify resources as simply commodities - clothing is a commodity which is ridiculous to lump with iron. And iron is not even like oil or molybdenum or semi-grade silicon or LOX or lanthanides.
Each of these has their own story. You can analyze coal or oil or solar or wind, in particular, and there seems nothing to learn from the game of piling them all into "commodity" or "technology" and expecting a story to fit. Sure you might gain the occasional insight - but nothing comparable to a serious analysis. Oil vs. solar for example each involve resource extraction, land use, pollution, questions of renewability, replacement rate, etc. In each of these dimensions you can make some interesting comparisons or contrasts.
But an overall narrative of technology vs. commodity? At the end of your essay nothing really emerges except perhaps "don't take guidance from such a superficial urge to categorize".
If you want to understand where things are headed, there is plenty of detail worth studying.
Enjoyed the read. The real obstacle isn’t whether renewables “work,” it’s that oil is so damn cheap, and so deeply embedded in our society/processes, that it’s become the easiest way to keep a complex economy running...which makes moving away from it structurally hard, even when people want to move to renewables. So frustrating.
The distinction is irrelevant. All energy sources require technology to extract energy from the commodity. Fossil fuel and hydro power generation technologies have been around a long time, and aren't likely to get much more efficient at the point of generation. Nuclear, solar, and wind generation technologies still have room to improve generating efficiency. Fossil fuel and nuclear fuels still have scope to increase in extraction efficiency. (Wind, solar, and hydro generate power where they're extracted, so the distinction there doesn't really matter.) But the longer a power source has been in use, the smaller the individual gains will generally be because there have been more attempts to improve.
Distributed production isn't always a net benefit. Generating power in individual vehicles is generally less efficient than generating power centrally and storing it in the vehicle, but has significant convenience benefits. At the scale of the national grid, 5GW power stations are still a "distributed" source; no one failure will cripple the system. It's easier to harden a single power station against the weather than it is to protect a million solar panels. But more finely distributed generation is more robust to transmission network failures.
Lastly, short and medium them price fluctuations aren't entirely cause by changes in the technology of extraction or generation. All phases of energy production, distribution, and finance are subject to political interference; major price changes can be purely the result of changes in policy halfway around the world.
It is puzzling that you would use materials from RMI which is simply an activist organization and their reports are always the same. This difference between technologies and commodities is not very useful and is mainly a way to buttress renewables instead of turning things like solar panels into commodities that are priced below market pricing from China. Much more interesting and useful would be to look at the work of Alvin Roth the recent winner of the Nobel in Economics who makes a much more meaningful distinction between commodity markets and matching markets and how to apply that thinking properly to the construction industry.
What if instead of "technology vs commodity" we expanded the analysis into "manufacture vs construction vs harvested vs extracted"? I feel like your book/work has shown that there are better learning curves for manufactures vs construction, and as you're showing now there appears to be some difference between agriculture and mining. I just mean "technology vs commodity" might be too simplistic, especially since as you point out, "technology" is kind of everything. But then it would still get to the point of "can things move from extraction to manufacture?" "Can things move from construction to manufacture?" How much does it help, and/or where does it break down? (Hydropower is construction, but if you get better at manufacturing them you still run out of places to put them. Windmills are manufactures but wake dynamics limit the useful supply. Etc.)
Yes, this. “Commodity” isn’t a natural kind. Consider:
- iron ore vs steel vs turbine tower
- silicon vs solar panel vs electricity
- fertilizer vs maize vs tortillas vs cattle feed vs beef
The simulations of production-curves-with-missing-data are showing >8 orders of magnitude on the x-axis: that seems unrealistic for looking at real-world datasets. Obviously someone somewhere collected the first bushel of grain, or mined the first kilogram of iron ore, but you wouldn't know its price; the best you'd know would be the annual production totals. For something like crops, those annual totals should grow pretty slowly over the 100-year stretch; e.g. for US crops, total annual production in 1900 is presumably >10% of what it is today (because population was ~1/3x), which means that the cumulative volume over that period should only span 3-4 orders of magnitude. It might be worth trying to simulate more realistic estimates of which x points you would expect to see.